Semiparametric Estimation of Fixed E¤ects Panel Data Models with Smooth

In this paper we consider the problem of estimating semiparametric \u85xed-e¤ects panel data models with smooth coe ¢ cients by local linear regression approach. We show that the proposed estimator has the usual nonparametric convergence rate and is asymptotically normally distributed under regular con-ditions. A modi\u85ed least-squared cross-validatory method is used to \u85nd the optimal bandwidth automatically. Moreover, we propose a test statistic for test-ing the null hypothesis of random e¤ects against \u85xed e¤ects for semiparametric panel data regression models with smooth coe ¢ cients. Monte Carlo simulations are used to study \u85nite sample performance of the proposed estimator and test